Short Fiction Ranking Introduction and Methodology

science fiction awards database

SF Novels   FH Novels   Novellas   Novelettes   Short stories

Introduction, April 2021:

The pages gathered here (in work) display ranked and chronological lists of the most signficiant SF/F/H short fiction and novels based on all the data compiled on this site: some 22,000 awards records (for works of fiction), some 15,000 citations (for books), and some 25,000 reprint records (for short fiction in anthologies), these data all weighted, and each item scored against its maximum possible total, as detailed below.

Scoring Methodology

The methodology for scoring and ranking short fiction is somewhat simpler than that for novels, but includes some steps that don't apply to novels. The methodology for novels will go into greater detail of these steps, with examples.

  1. Collecting: Awards data, citations (for short fiction these are so far limited to all-time best polls), and records of reprints (of stories in the some 1400 anthologies profiled on this site) have been compiled. A simple tally of the number of such references for each story is the first approximation of the sfadb score.
  2. Weighting:
    • Different awards, citation sources, and anthologies are weighted by perceived authority and significance, with points assigned to each story title correspondingly, and summed by title.
      • For awards, established awards that are juried are weighted most highly; awards limited to specific themes or author nationalities, and those most open to popular voting, the least.
      • For reprints, those anthologies with the most explicit aims to gather the best or most significant stories of all time, or of a particular era, and weighted most highly; those limited to specific themes rather less; while original anthologies are given only token weights.
    • In addition, the total number of additional reprints (including both anthologies and single-author collections) beyond those in the profiled anthologies is factored into the weighting. (These counts are extracted from the Locus Index to Science Fiction, or for recent stories from This is to account for stories that are widely known from author collections and perhaps therefore not widely anthologized, e.g. Bradbury's "A Sound of Thunder" and several Heinlein stories.
    • The result is the second approximation of sfadb score.
  3. Scoping: For short fiction, this is assigning each story to be a novella, novelette, or short story, depending on length. In practice story lengths indicated in the Locus Index to Science Fiction have been adopted, with a handful of manual corrections. For more recent stories, award categories, or length indications at, have been used.
  4. Scaling: The maximum number of points each work *might* have earned – given year of publication – is calculated.
    • For awards nominations and wins, each story is not scored against every possible award given for its year, but rather only against the principle awards that presumably any story would eligible for -- e.g. Hugo, Nebula, Locus, Sturgeon, and so on. This demotes awards of limited scope (theme, origin of author or publication, etc.) to be in effect 'bonus' awards that increase a story's score asymptotically (the story's actual points and possible points are both increased by the weight of the award), but does not penalize all the stories that were not nominated for the award. The distinction is shown on the ranking page by those award abbreviations shown before the vertical bar, |, and those after.
    • For anthology reprints, each story is not scored against every possible anthology it could theoretically have been reprinted in, but rather against groups of anthologies that, by inspection, are more-or-less mutually exclusive (e.g. some editors rarely use the same story twice, across dozens of anthologies). The effect is to reduce the 1400 weighted anthologies down to three dozen or so scoring groups, not all of which, of course, apply to every story. (Without this approach, actual percentage scores would be minuscule, even for the highest ranking stories.)
  5. Scoring: Titles are scored by the percentage of actual points to possible points. This is the third approximation of sfadb score.
  6. Proportioning: A cap is placed on the proportion awards points contribute to a novel’s or story’s sfadb score. This is to prevent recent works that do well in awards from jumping to the top of the rankings; in effect requiring a novel or story to accumulate, over the decade or two following its initial publication, significant citations or reprints before it moves into the top 100 of any ranking. The degree to this dampening, or proportioning, is adjusted to put a work at around rank #100 in its category should it win every major award it's eligible for. (It could rank higher with nominations or wins of 'marginal' or 'bonus' awards.) This is the final sfadb score.
  7. Ranking: Works are sorted by their final sfadb scores on the five ranking pages for SF novel, FH novels, novellas, novelettes, and short stories. Descriptions and references are provided for the top 20 on each page, initially.
  8. Tracking: Grids of abbreviations for possible and actual award, citation, and reprint references, with links on each, are displayed for each title on the ranking pages.
  9. Profiling: Extended results from all five categories are profiled across time on the Timeline page.

Copyright 2012 - 2021 by Mark R. Kelly and the Locus Science Fiction Foundation. All rights reserved.